BIOMEDICAL AND ENVIRONMESTAL SCIENCES 10. 415.435( 1997)

Environmental Implications of ExcessiveSelenium: A Review

A. DENNIS LEMLY

United States Forest Service, Southern Research Station, ColdwaterFisheries Research C’nit , Department of Fisheries and Wildlife

Sciences, Virginia Tech University, Blacksburg, VA24061-0321, U S A

Selenium is a naturally occurring trace element that is nutritionally required in smallamounts but it can become toxic at concentrations only twice those required. The narrow mar-gin between beneficial and harmful levels has important implications for human activities thatincrease the amount of selenium in the environment. Two of these activities, disposal of fossilfuel wastes and agricultural irrigation of arid, seleniferous soils, have poisoned fish andwildlife, and threatened public health at several locations in the United States. Research stud-ies of these episodes have generated a data base that clearly illustrates the environmental hazardof excessive selenium. It is strongly bioaccumulated by aquatic organisms and even slight in-creases in waterborne concentrations can quickly result in toxic effects such as deformed em-bryos and reproductive failure in wildlife. The selenium data base has been very beneficial indeveloping hazard assessment procedures and establishing environmentally sound water qualitycriteria. The two faces of selenium. required nutrient and potent toxin, make it a particularly

~ important trace element in the health of both animals and man. Because of this paradox, envi-ronmental selenium in relation to agriculture, fisheries, and wildlife will continue to raise im-portant land and water-management issues for decades to come. If these issues are dealt withusing prudence and the available environmental selenium data base, adverse impacts to naturalresources an&public health can be avoided.

INTRODUCTION

Selenium is a trace element that is normally present in surface waters at concen-trations of about 0.1-O. 3,ug/l (parts-per-billion; Lemly, 1985a). In slightly greateramounts, i. e. , l-5pg/l, it can bioaccumulate in aquatic food chains and become aconcentrated dietary source of selenium that is highly toxic to fish and wildlife (Lemlyand Smith, 1987; Lemly, 1993a). Dietary selenium is passed from parents to off-spring in the eggs, where it can be teratogenic to developing embryos and cause com-plete reproductive failure (Gillespie and Baumann, 1986; Heinz et al. , 1987, 1989;Coyle et al. , 1993 ; Lemly, 1993b). This scenario of poisoning has occurred inreservoirs contaminated by selenium leached from fly-ash at coal-fired electric generat-ing stations in the eastern U. S. (Garrett and Inmann, 1984; Lemly, 1985b). andin wetlands used to dispose subsurface agricultural irrigation drainage in the westernU .S . (Lemly , et al., 1993; Lemly 1994a, 199413). Public health may also bethreatened if humans consume selenium-contaminated fish and wildlife (Fan et al.,1988).

0895-3988/97CN 11-2816Copyright 0 1997 by CAF’M

415

416 A. DENNIS LEMLY

Because of these and other episodes of contamination, the role and importance ofselenium as an environmental pollutant has gained widespread attention among re-search scientists, natural resource managers, and regulatory agencies in the U. S.during the past decade. Although the basic toxicological symptoms and paradox of se-lenium (nutritionally required in small amounts by both animals and humans, buthighly toxic in slightly greater amounts) have been known for many years (Draizeand Beath, 1935; El l is et al., 1937; Rosenfeld and Beath, 1946; Hartley andGrant, 1961), it was not until the late 1970s and early 1980s that the potential forwidespread contamination of aquatic ecosystems due to human activities was recog-nized ( Andren et al. , 1975; Cherry and Guthrie, 1977; Evans et al., 1980; Na-tional Research Council, 1980a; Braunstein et al. , 1981). In fact, as recently as1970 selenium was being called the “unknown pollutant” with respect to what wasknown about its cycling and toxicity in the aquatic environment (Copeland, 1970).Since that time much has been learned about the environmental implications of exces-sive selenium. This article is a concise review that gives a current perspective on ma-jor sources of selenium, impacts and threats to wildlife and humans, ways to assessenvironmental hazard, and criteria for protecting environmental quality and publichealth.

MAJOR SOURCES OF SELENIUM CONTAMINATION

Two factors stand apart as the major human-related causes of selenium mobiliza-tion and introduction into the environment. First is the procurement, processing, andcombustion of fossil fuels. Selenium is an important trace element present in coal,crude oil, oil shale, coal conversion materials (liquefaction oils and synthetic gases),and their waste by-products (Pillay et al. , 1969 ; American Petroleum Institute,1978 ; Fruchter and Petersen, 1979; Schlinger and Richter, 1980; Clark et aL . ,1980; U. S.EPA, 1980; Cowser and Richmond, 1980; Nystrom and Post, 1982).Rain can leach selenium from coal and oil-shale mining, preparation, and storagesites, where it may enter down-gradient streams and reservoirs through precipitationrunoff (Davis and Boegly, 1981; Heaton et al., 1982, Jones, 1990). More impor-tantly, selenium is highly concentrated in the mineral fraction (fly-ash and bottomash) remaining after coal is burned ( Kaakinen et al. , 1975 ; Klein et al. , 1975).Over 70 million tons of coal fly-ash are produced annually in the U. S. and most of itis disposed by dumping into wet-slurry or dry ash basins (Murtha et al. , 1983). Se-lenium-laden leachate and overflow from these basins often makes its way into rivers,streams, and impoundments. Selenium concentrations can rapidly increase in fish andaquatic organisms in the receiving water, ultimately resulting in tissue damage, re-productive failure, and elimination of entire fish communities (Cumbie and VanHorn, 1978; Garrett and Inmann, 1984; Lemly, 1985a, 1985b; Sorensen, 1986).

The second major factor is the irrigation of seleniferous soils for crop productionin arid and semi-arid regions. For example, deposits of Cretaceous marine shales haveweathered to produce high selenium soils in many areas of the western U. S. , notablythe San Joaquin Valley of California and certain parts of Wyoming, Colorado, Neva-da, North and South Dakota, Montana, New Mexico, Arizona, Utah, Nebraska,and Kansas (Kubota, 1980; Tanji et al., 1986; Allen and Wilson, 1990; Presserand Ohlendorf , 1987; Presser et al., 1990; Severson et al., 1991) . These areasusually require substantial irrigation for agricultural crop production. Excessive

SELENIUM IN THE ENVIRONMENT 417

amounts of irrigation water are necessary to leach away salts and prevent salinizationof soils. This can lead to the production of subsurface drainage that may be highlycontaminated with dissolved selenium and other soil trace elements that have beenleached from the soil along with salts ( Presser and Barnes, 1985; Saiki, 1986a;Summers and Anderson, 1986; Fuji, 1988; Deverel et al., 1989). Selenium inagricultural irrigation drainwater was responsible for massive poisoning of fish andwildlife at Kesterson National Wildlife Refuge, California in the early to mid-1980s(Marshall, 1985; Hoffman et al., 1986; Saiki, 1986a, 1986b; Saiki and Lowe,1987; Ohlendorf et al., 1986, 1988a, 1988b). Subsequent studies have shown irri-gation-related selenium contamination to be a threat to aquatic systems and wildliferefuges in many western states (U. S. Fish and Wildlife Service, 1986; Summers andAnderson, 1986; Ohlendorf et al. , 1987; Sylvester et al. , 1991).

AQUATIC CYCLING AND BIOACCUMULATION

The Selenium Cycle

Four primary pathways exist for selenium in aquatic systems: (i) it can be ab-sorbed or ingested by organisms, (ii) it can bind or complex with particulate matter,(iii) it can remain free in solution, or (iv) it can be released to the atmospherethrough volatilization. Over time, most selenium is either taken up by organisms orbound to particulate matter. Through deposition of biologically incorporated seleniumand settling of particulate matter (sedimentation), most of it usually accumulates inthe top layer of sediment and detritus. However, sediments are only a temporaryrepository for selenium because there are numerous biological, chemical, and physicalprocesses that can move it out of sediments as well. Aquatic systems are very dynamicand selenium can be cycled from sediments into biota and remain at elevated levels foryears after waterborne inputs of selenium are stopped.

Immobilization and Removal Processes

Selenium can be removed from solution and sequestered in sediments through thenatural processes of chemical and microbial reduction of the selenate form (Se VI) tothe selenite form (Se IV), followed by absorption (binding and complexation) ontoclay and the organic carbon phase of particulates, reaction with iron species, and co-precipitation and settling. Regardless of the route, once selenium is in the sediments,further chemical and microbial reduction may occur, resulting in insoluble organic,mineral, elemental, or adsorbed selenium. Most selenium in animal and plant tissuesis eventually deposited as detritus and, over time , isolated through the process of sed-imentation. In total, immobilization processes effectively remove selenium from thesoluble pool, especially in slow moving or still-water habitats and wetlands. Ninetypercent of the total selenium in an aquatic system may be in the upper few centimetersof sediment and overlying detritus (Lemly and Smith, 1987).

Mobilization Processes

Selenium in sediments is particularly important to long-term habitat quality be-cause mechanisms present in most aquatic systems effectively mobilize this selenium

418 A. DENNIS LEMLY

into food chains and thereby cause long-term dietary exposure of fish and wildlife. Se-lenium is made available for biological uptake by four chemical and/or biological pro-cesses. The first is the oxidation and methylation of inorganic and organic selenium byplant roots and microorganisms (oxidation is the conversion of selenium in the reducedorganic, elemental, or selenite forms to the selenite or selenate forms; methylation isthe conversion of selenium to an organic form containing one or more methyl groups,which usually results in a volatile form). The second process is the biological mixingand associated oxidation of sediments that results from the burrowing of benthic in-vertebrates and feeding activities of fish and wildlife. The third process consists ofphysical perturbation and chemical oxidation associated with water circulation andmixing ( current, wind, stratification, precipitation, and upwelling). Finally, sedi-ments may be oxidized by plant photosynthesis. Two additional pathways provide fordirect movement of selenium from sediments into food chains, even when the surfacewater does not contain elevated concentrations of the element. These pathways areuptake of selenium by rooted plants and uptake by bottom-dwelling invertebrates anddetrital-feeding fish and wildlife. These two pathways may be the most important inthe long-term cycling of potentially toxic concentrations of selenium. Thus, rootedplants and the detrital food pathway can continue to be highly contaminated and ex-pose fish and wildlife through dietary routes even though concentrations of selenium inwater are low ( Lemly and Smith, 1987) Some of the sedimentary and waterborneselenium may be converted to volatile organic forms (e. g. , dimethyl selenide) by mi-crobial activity in the water and sediments, and subsequently released (degassed) intothe atmosphere through the water-air interface or through direct release by plants.Volatilization is an important mechanism by which selenium can be removed fromaquatic systems and, thereby, help alleviate the threat of selenium toxicosis to fishand wildlife (Karlson and Frankenberger, 1989, 1990).

Bioaccumulation

The major environmental implications of excessive selenium are associated withits propensity to bioaccumulate in aquatic food chains and, thereby, contaminate thediet of fish and wildlife and, in some cases, humans. Aquatic organisms can accumu-late this element to concentrations one or more orders of magnitude greater than theconcentrations in their water or food (Fig. 1) . The reason for this bioaccumulationmay be that selenium is chemically similar to sulfur, and it sometimes is an essentialmicronutrient for animals. Over evolutionary time, aquatic animals may have evolvedmechanisms to retain and accumulate selenium under conditions of scarcity. Whateverthe reason, bioaccumulation has important implications for toxic effects. For exam-ple, where fish have experienced chronic toxicity, selenium in the water has beenconcentrated from 100 to more than 30,000 times, depending on the species and tis-sue sampled. Selenium accumulation in the organisms eaten by fish and wildlife isusually the major pathway leading to toxicity. Biomagnification of selenium (the ac-cumulation of progressively greater concentrations by successive trophic levels of afood chain) usually ranges from 2 to 6 times between the primary producers (algaeand plants) and the lower consumers (invertebrates and forage fish). For example,fish that eat contaminated plankton or benthic invertebrates may accumulate seleniumto concentrations that are 4 times those of their food, which in turn, could contain500 times the selenium concentration in the water. The food chain biomagnification

SELENIUM IN THE ENVIRONMEXT 419

factor would then be 4 and the total bioconcentration factor for the fish would be2,000 (i. e. , 4 X 500). These relationships are important in natural systems becausethey can cause top-level consumers, such as predatory fish, birds, and mammals, toreceive toxic selenium levels in the diet even though the concentration in water is low(< lOpg/liter) Moreover, the risk of toxicity through the detrital food pathway willconti-nue despite a loss of selenium from the water column, as long as contaminatedsediments are present (Lemly and Smith, 1987).

FIG. 1. A typical scenario for bioaccumulation of selenium in aquatic food chains.In this example the bioaccumulation factor for fish relative to water is 3,750 for

the planktonic food pathway (750 X 5) , and 5,250 for the benthic food pathway(350 X 3 X 5). The food-chain biomagnification factor for the invertebrate-to-fishtrophic step is 5 for both pathways.

Role of Habitat Variability

The processes regulating selenium cycling are very similar in all aquatic habitatsand wetlands. However, the relative importance of each process may vary from onelocation to another depending on hydrologic factors that affect food chain bioaccumu-lation and exposure of wildlife. For example, in fast flowing waters, fine organic sed-iments ‘such as those produced by the deposition and decay of particulate matter andplant and animal tissue may be rare because they are continually flushed out of thesystem. In such waters there is little opportunity for a contaminated surface layer ofsediment to develop. Moreover, rooted plants and benthic invertebrates are oftenscarce in this habitat type. The benthic-detrital component of the system and the as-sociated food pathways thus play a smaller role in the selenium cycle in lotic (flowing)waters than in lentic (slow water) habitats such as reservoirs and wetlands. Systemsthat tend to accumulate selenium most efficiently are shallow wetlands and marshes,

420 A. DENNIS LEMLY

and reservoirs with low flushing rates. In these systems biological productivity is of-ten high and selenium may be readily trapped through immobilization processes orthrough direct uptake by organisms. Sediments often build up a high selenium con-centration that is remobilized gradually and continually through detrital and plankton-ic organisms. These habitats are also typically some of the most important !peding andbreeding habitats for fish and wildlife, especially waterfowl and shorebirds. Severalhabitat types often occur together in one aquatic system or wetland complex. For ex-ample, rivers may have fast flowing waters, slow moving pools, and shallow backwa-ter marshes all within a few hundred meters of each other. The degree of fish andwildlife exposure to selenium varies among habitats according to the intensity of use,type of use ( e. g . , feeding vs. resting), and the relative contributions of the variousprocesses that regulate selenium cycling and bioaccumulation in food chains (Lemlyand Smith, 1987).

THREATS TO ECOSYSTEMS

Case Example 1. Toxic Impacts to Fish

Belews Lake, North Carolina was contaminated by selenium in wastewater re-leased from a coal-fired electric generating facility. From 1974 through 1985, waterwas withdrawn from the lake and mixed with bottom ash from the coal burners andfly-ash collected by electrostatic precipitators. This slurry was pumped from the pow-er plant and discharged into a 142 hectare ash basin, where suspended solids were col-lected by gravitational settling. Runoff water from the coal storage area and powerplant site was collected by sump units and also pumped into the ash basin. Selenium-laden (150-2OOpg/l) return flows from the ash basin entered the west side of BelewsLake through an ash sluice water canal.

Selenium bioaccumulated in aquatic food chains and caused severe reproductivefailure and teratogenic deformities in fish (Cumbie and Van Horn, 1978; Lemly,1985a, 1985b, 1993b). Congenital malformations consisted of missing fins, protrud-ing eyes, and grossly deformed spines and heads (Fig. 2). Concentrations of seleniumin the lake water averaged only lO,ug/l, but were accumulated from 519 times (peri-phyton) to 3,975 times (visceral tissue of fish) in the biota. The pattern and degreeof accumulation was essentially complete within 2 years after the initial operation ofthe power plant, and persisted throughout the period of selenium discharge into thelake ( 1974-1985). Highest concentrations of selenium were found in fish, followedby benthic macroinvertebrates, plankton, and periphyton. The planktonic and detri-tal food pathways exposed fishes to potential dietary concentrations of selenium thatwere some 770 and 510-l) 395 times the waterborne exposure, respectively. Of the20 species of fish originally present in the reservoir, 16 were eliminated (which in-cluded all of the important gamefish species) through a combination of dietary toxicityand reproductive failure (Fig. 3). Two species were rendered sterile but persisted asadults and one additional species was eliminated but managed to partially recolonizefrom a relatively uncontaminated headwater area. Only one of the original residentspecies, the selenium-tolerant mosquitofish ( Gambusia affiinis > survived, along withtwo introduced cyprinids. The severe toxic impacts in Belews Lake occurred eventhough concentrations of waterborne selenium were only lo-20 times those in nearbyuncontaminated reservoirs; the flora and fauna contained only about lo-15 times as

SELEKIUM IN THE ENVIRONMENT 421

FIG. 2. An example of selenium-induced teratogenic deformities infish. These are mosquitofish ( Gambusia affinis) collected fromBelews Lake, North Carolina during the period of selenium con-tamination. The top two individuals are deformed and have severekyphosis (convex curvature of the thoracic region of the spine,Arrow 1) and lordosis (concave curvature of the lumbar region ofthe spine, Arrow 2). They also are missing or have vestigialpelvic fins (Arrow 3 ) The bottom individual is normal.

much selenium.The findings from Belews Lake serve as a clear illustration of how selenium can

rapidly impact fish populations. Moreover, this case example highlights the fact thatselenium can accumulate and be biologically magnified to toxic levels even though wa-terborne concentrations are in the low microgram per liter range. Were it not forbioaccumulation, these waterborne concentrations would pose little threat to aquaticlife. Information from the field studies of selenium toxicity in Belews Lake was in-strumental in the U. S. Environmental Protection Agency’s decision to lower the na-tional water quality criterion for selenium from 35pg/l to 5pg/l (U. S. EPA, 1987).

In response to concerns about the fishery problems in Belews Lake the electric u-tility company changed its ash disposal practices. This involved switching to a dry-ashhandling system that disposed the waste in a landfill rather than a wet-basin. By mid1986, selenium-laden wastewater no longer entered the lake (North Carolina Depart-ment of Natural Resources and Community Development, 1986).

Follow-up studies were conducted in 1996 to assess recovery of the ecosystem inBelews Lake (Lemly, 1997a). Selenium concentrations and associated impacts to fishwere measured and compared to pre-1986 conditions to determine how much changeoccurred during the decade since selenium inputs stopped. Findings were also exam-ined using a hazard assessment protocol ( Lemly, 1995) to determine if ecosystem-lev-

422 .A. DENNIS LEIMLY

Ictalurtdae

Cyprinidaa

Poaciliidae -

-_ Centrarchidae I-Catortomidae

- Percichthyidae

\-Ctupeidao

Percidao

t 1973,\,‘I/., , , , , ,

75f t

77 79 81 83

Unit I Unit 2

-e-_)

FIG. 3. The pattern of poisoning of fish in Belews Lake, North Carolina due to selenium conrami-nation. Selenium-laden wastewater (150-2OOpg/l) was discharged into the lake begining in 1974and within three years most of the fish ( 16 species) had been eliminated due to dietary toxicity andreproductive failure.

el hazards to fish and aquatic birds had changed as well. Results showed that water-borne selenium fell from a peak of 2Opg/l before 1986, to < lpg/l in 1996; concen-trations in biota were 85-95 % lower in 1996. Hazard ratings indicated that high haz-ard existed prior to 1986 and that moderate hazard was still present in 1996, primari-ly due to selenium in the sediment-detrital food pathway. Concentrations of seleniumin sediments fell by about 65-75 96 during the period but remained sufficiently elevat-ed (l-4pg/g) to contaminate benthic food organisms of fish and aquatic birds. Fieldevidence confirmed the validity of the high hazard ratings. Developmental abnormali-ties in young fish persisted in 1996. This indicated that selenium-induced teratogene-sis and reproductive impairment were still occurring. Moreover, the concentrations ofselenium in benthic food organisms were sufficient to cause mortality in young bluegilland other centrarchids because of Winter Stress Syndrome (Lemly, 1993~ 1996a).At the ecosystem level, recovery in Belews Lake has been slow. Toxic effects werestill evident ten years after selenium inputs were stopped. The sediment-associated se-lenium will likely continue to be a significant hazard to fish and aquatic birds for yearsto come in this aquatic system.

Case Example 2. Toxic Impacts to Wildlife

In 1985, subsurface irrigation drainage was implicated as the cause of death anddeformities in thousands of waterfowl and shorebirds at Kesterson National WildlifeRefuge in California (Ohlendorf et al. , 1986). Naturally occurring trace elementsand salts were leached from soils on the west side of the San Joaquin Valley and car-ried to the refuge in irrigation return flows that were used for wetland management(Zahm, 1986). One of the trace elements, selenium, bioaccumulated in aquatic foodchains and contaminated 500 hectares of shallow marshes. Elevated selenium wasfound in every animal group coming into contact with these wetlands- f r o m f i s hand birds to insects, frogs, snakes, and mammals (Saiki and Lowe, 1987; Clark,1987; Ohlendorf et al. , 1988a). Congenital malformations in young waterbirds were

SELENIUM IN THE ENVIRONMENT 423

severe, and included missing eyes and feet, protruding brains, and grossly deformedbeaks, legs, and wings (Ohlendorf et ul., 1986, 1988b; Hoffman et (IL., 1988).Several species of fish were eliminated due to the combined effects of high salinity, el-evated selenium, and other contaminants (Saiki et al., 1992). A high frequency( 30 % ) of stillbirths occurred in the single remaining species, mosquitofish GumbusiuuffiniJ (Saiki et ul., 1991). Laboratory studies conducted by the U. S. Fish andWildlife Service confirmed the field assessment that selenium in irrigation drainagewas the cause of the fish and wildlife problem (Lemly et al. , 1993). The “poisoned”refuge became highly publicized and sparked a great deal of political and scientific con-troversy (Marshall, 1985; Popkin, 1986; Harris, 1991).

The findings at Kesterson National Wildlife Refuge led to a new awareness of theenvironmental hazards posed by selenium in agricultural irrigation drainage. In 1986,the U.S. Department of the Interior, the Federal steward of more than 400 irriga-tion-drainage facilities and 200 wildlife refuges in the western states (U. S. Bureau ofReclamation, 1981)) established a multi-agency program to investigate irrigation-re-lated drainwater problems. This program conducted screening-level assessments in 13states, including 20 national wildlife refuges (Table 1) . The western San JoaquinValley and Kesterson National Wildlife Refuge were used as models for identifyingand prioritizing potential study areas based on the occurrence of conditions known tocontribute to drainwater problems. Samples of water, sediment, and biota (inverte-brates, whole-fish, bird liver, bird eggs) were analyzed for a variety of trace ele-ments, heavy metals, and pesticides, and the results were compared to concentrationsknown to be toxic to fish and wildlife in experimental studies. Geological and hydro-logical studies were’ conducted and, where possible, observations were made to docu-ment the occurrence of deformed embryos and hatchlings, which is a biomarker forselenium poisoning in birds (Hoffman and Heinz, 1988).

By 1992 it was known that eleven of the sixteen study areas where biologicalsamples had been taken were seriously contaminated by selenium. The concentrationspresent at these eleven sites exceeded toxicity thresholds for fish and wildlife (Presseret ul., 1994) . These study areas are spread across nine states (Fig. 4) Overt seleni-um toxicosis -i. e. , deformities in bird embryos and hatchlings-was found in fivestates; California, Utah, Wyoming, Nevada, and Montana (Fig. 4, Table 1). Insome cases, these teratogenic effects occurred even though the waterborne concentra-tions of selenium were below those recommended by the U. S EPA for the protectionof aquatic life (Lemly et al. , 1993).

The biogeochemical conditions leading to the production of subsurface irrigationdrainage, culminating in death and deformities in wildlife, have been termed the“Kesterson Effect” (Presser, 1994). The Kesterson Effect is prevalent throughoutthe western United States and consists of these key conditions: (i) a marine sedimen-tary basin that contains Cretaceous soils, which usually have relatively high concen-trations of selenium; (ii) alkaline, oxidized soils that promote the formation of water-soluble forms of selenium and other trace elements; (iii) a dry climate in which evap-oration greatly exceeds precipitation, leading to salt buildup in soils; (iv) subsurfacelayers of clay that impede downward movement of irrigation water and cause water-logging of the crop root zone; and ( v) subsurface drainage, by natural gradient orburied tile drainage networks, into migratory bird refuges or other wetlands.

424 A. DESNIS LEMLY

TABLE 1

Study Areas and National Wildlife Refuges Investigated m the U.S. Department of the Interior’sIrrigation Drainage Program ( Lemly er al. , 1993)

State and Study Area National Wildife Refuge (NWR)

- -OregonMalheurb

Oregon/CaliforniaMalheur NWR

CaliforniaKlamath Basin

Sacramento Complex

Tulare Lake Bed’

Salton ScabCalifornia/Arizona

Lower Colorado River

Nevada

Utah

Montana

Colorado

Wyoming

Lower Klamath NWR

Sacramento NWRDelevan NWRColusa NWRSutter NWRKern NWRPixley NWRTule Lake NWRSalton Sea NWR

Havasu NWRCibola NWRImperial NWR

Stillwater’

Middle Green River’

Sun River’Milk River Basin

Gunnison River BasinbPine River

Kendrick Project’Riverton Projectb

Stillwater NWR

Ouray NWR

Benton Lake NWR-

--

Bowdoin NWR-

South DakotaBelle Fourche ProjectbAngostura Project

KansasMiddle Arkansas Riverb

TexasLower Rio Grande Valley

New MexicoMiddle Rio Grande Valley

IdahoAmerican Falls Reservoir

--

-

Laguna Atascosa NWR

Basque de1 Apache NWR

Minidaka NWR

’ Study areas where overt symptoms of selenium toxicosis (deformities) were found in young migratory birds.’ Study areas where toxicity is predicted based on concentrations of selenium found in fish and bird tissues.

SELENIUM IN THE ENVIRONMENT 425

F1c.4. Locations in the western United States where selenium in subsur-face agricultural irrigation drainage has poisoned fish and wildlife. 1 = Mal-hew National Wildlife Refuge (NWR) ; 2 = Stillwater NWR; 3 = TulareLake Bed Area: 4 = Salton Sea Area; 5 = Benton Lake NWR; 6 = BelleFourche Reclamation Project; 7 = Bowdoin NWR; 8 = Riverton Reclama-tion Project; 9 = Ouray NWR; 10 = Gunnison River Basin: 11 = MiddleArkansas River. Teratogenic deformities associated with selenium bioaccu-.mulation in young birds were found at locations 2, 3. 5. 7, and 9.

The field studies conducted by U. S. Department of the Interior indicate that thetoxic threat of selenium-laden irrigation drainage to wetlands, fish, and wildlife is notrestricted to Kesterson National Wildlife Refuge, the San Joaquin Valley, or the Stateof California. Contamination has proven to be pervasive throughout the western states(Fig. 4 > and threatens waterfowl populations in the Central and Pacific flyways(Presser et al. , 1994; Skorupa et al., in press). Managers of wetlands in the west-ern U. S. and elsewhere must recognize selenium-laden irrigation drainage as a threatwith the potential to affect wildlife populations on an international scale.

THREATS TO HUMAN HEALTH

Excessive selenium in the environment can threaten public health in the sameway it threatens fish and wildlife, i.e. , by accumulating in potential food items. Insome locations soils that are naturally high in selenium may cause selenium concentra-tions in agricultural crops to be excessive, leading to chronic poisoning in humans

426 A. DENNIS LEMLY

(Yang et al., 1983). For situations in which human activities have led to contami-nated aquatic systems, public health risks are associated with consumption of seleni-um-contaminated fish and wildlife. The U. S. National Academy of Sciences and theU.S. Environmental Protection Agency recommend that a daily intake of 50.2OOvgselenium should be adequate for nutritional requirements in adults (National Research‘Counc i l , 1980b; U.S. EPA, 1984). Th e dose necessary to cause chronic selenosis inhumans is not well defined but the threshold for toxicity appears to lie somewhere inthe range of GO-9OOpg per day (Yang et al., 1989a, 1989b; Longnecker et al.,1991) . Thus, the margin of safety for selenium, i. e. , the difference between the di-etary levels that are beneficial and those that are possibly harmful, is small-a factorranging from about 5 to 18.

In Case Example 2, public health agencies recognized the possibility that seleni-um poisoning could occur in humans that ate fish and wild game taken from contami-nated sites. Federal land management authorities erected warning signs at these loca-tions to alert the public to the health risks associated with selenium contamination andguard against consumption of contaminated food (Fig. 5) . Dietary guidelines formu-lated by the California Department of Health stipulated that adults should eat no morethan one meal (4 ounces) of fish or wild game per 2 weeks (Moore et al. , 1990).Children less than 15 years of age and women of childbearing age were advised not toeat any fish or wild game due to concerns about potential reproductive and develop-mental effects. Advisories were issued when the selenium concentrations reached orexceeded 2 pg/g wet weight (Fan et al. , 1988 > . The rationale for this thresholdconcentration was as follows: By simple calculation, fish or game containing 1 ,ug/gselenium contributes 113 pg per 4 oz serving. At 2 ,ug/g the single serving would con-tribute 226 pg, which exceeds the upper limit of the recommended daily intake

GOVERNMENT PROPERTYNO TRESPASSING

Potential health risk fromselenium contamination

F I G. 5. An example of the signs used toalert the public of the health risks from se-lenium contamination at Federally managedwetlands in the western United States.

SELENIUM IN THE ENVIRONMENT 427

(200 pg) This exceedance would be due to a single meal, not including other dailyintake. Therefore, a concentration of 2 pg/g was designated as the action level for is-suing consumption advisories. A complete ban on consumption was issued and postedat sites where concentrations of selenium in fish or bird tissues exceeded 5 ,ug/g wetweig_ht (a concentration of 5 ,ng/g would contribute 565 I-(g per serving, which is inexcess of the 500 pg/day chronic toxicity threshold suggested by Levander, 1987).

The risks to public health in Case Example 2 prompted the first widely publicizedconsumption advisories and bans due to excessive environmental selenium in the Unit-ed States. These advisories were issued in 1984-85 (Fan et al., 1988). Since thattime, several other advisories have been issued for selenium contaminated reservoirsand wetlands. For example, the fishery of Belews Lake (Case Example 1) began torecover after 1986 as selenium levels fell, but the concentrations in fish tissues werestill great enough (3-8pg/g) to pose a threat to human health. This threat was recog-nized and advisories were issued begining in 1989 (Lemly, 1997a). The stipulationsof those advisories paralleled the ones given above. Warnings have also been posted atseveral wildlife refuges contaminated by seleniferous irrigation drainage in the westernU. S (e.g., Stephens et al., 1992; Butler et al., 1994). It is now widely recog-nized in the U. S. that excessive selenium in the environment can, and does, pose athreat to human health as well as fish and wildife. In regard to the two case examplesdiscussed in this paper, timely intervention by public health agencies prevented whatcould have otherwise developed into episodes of human poisoning. Other countrieswould be well advised to take note of these examples.

HAZARD ASSESSMENT AND ENVIRONMENTAL QUALITY CRITERIA

It is important to monitor and evaluate environmental conditions as a way of pre-dicting, detecting, and, hopefully, avoiding potential selenium problems before theyoccur. A hazard assessment protocol (Protocol) for selenium has recently been devel-oped and is now available for use in this regard (Lemly, 1995, 1997b). The Protocolcharacterizes hazard in terms of the potential for food-chain bioaccumulation and re-productive impairment in fish and aquatic birds, which are the most sensitive biologi-cal responses for estimating ecosystem-level impacts of selenium contamination. Fivedegrees of hazard are possible depending on where the highest concentrations of sele-nium measured in environmental samples fall on the corresponding hazard profile giv-en in the Protocol (Fig. 6) . The degree of hazard is given a numerical score: 5 = highhazard, 4 = moderate hazard, 3 = low hazard, 2 = minimal hazard, and 1~ no identi-fiable hazard. A separate hazard score is given to each of five ecosystem components;water, sediments, benthic macroinvertebrates, fish eggs, and aquatic bird eggs. Afinal hazard characterization is determined by adding individual scores and comparingthe total to the following evaluation criteria: 5 = no hazard, 6-8 = minimal hazard, 9-11 = low hazard, 12-15 = moderate hazard, 16-25 = high hazard ( Fig. 6, Table 2 ;Lemly, 1995, 1997b).

Selenium hazards to fish can vary seasonally because of a condition known asWinter Stress Syndrome (WSS; Lemly, 1993~ 1996a, 1997~). WSS is severe lipiddepletion brought on by external stressors in combination with normal reductions infeeding and activity during cold weather. Fish can develop this syndrome in responseto chemical stressors such as water pollutants, or biological stressors such as para-sites. Substantial mortality can result, potentially changing year-class strength and

428 A. DENNIS LEIMLY

population structure of the affected species, and altering community-level ecologicalinteractions. Selenium hazards should be evaluated in the context of seasonal metabol-ic changes that normally occur in test organisms. WSS could be an important cause ofmortality in many circumstances. Wastewater discharges containing selenium maypose a greater toxic threat to fish during winter than at other times of the year (Lem-ly;- 1997d). A comprehensive protocol for aquatic hazard assessment should includetesting for WSS. Because of the possibility of WSS, conservative estimates of hazardare necessary and the role of season in aquatic hazard assessment must be accountedfor.

TABLE 2

Example Data Sets for Aquatic Hazard Assessment of Selenium Followmg the Protocol Method(Lemly, 1995)

Site and Environmental Selenium Evaluation byComponent Concentration” Component Totals for the Site

Haaardb Score Score HazardWetland X

Water

Sediments

Invertebrates

Fish eggs

Bird eggs

Reservoir X

Water

Sediments

Invertebrates

Fish eggs

Bird eggs

River X

Water

Sediments

Invertebrates

Fish eggs

Bird eggs

< 1-3

0.7-l

1-3

2-4

2-7

9-93 High

7-41 High

12-72 High

75-120 High

12-120 High

3-4 Moderate

0.6-3 Low

3-33 High

8-27 High

1-17 Moderate

Low

None

Minimal

Minimal

Low

Low

3

1

2

2

311 11

5

5

5

5

525 25

4

3

5

54

21 21 High

High

“Selenium concentrations in pg/l (parts per billion) for water; mg/g (parts per million) dry weight for sedi-ments, invertebrates, and eggs.

b Hazard ratings were determined by comparing selenium concentrations to hazard profiles given in Fig. 6.

In addition to the Protocol, which predicts environmental hazard, there are alsorecent guidelines for evaluating the toxicological significance of selenium residues in

SELENIUM IN THE ENVIRONMEST 429

aquatic organisms and recommendations for water quality criteria to protect fish andwildlife (Lemly, 1993a, 1996b). Diagnostic selenium concentrations are available forwater, food-chain, predatory fish (consuming fish or invertebrate prey), and aquaticbirds. Waterborne selenium concentrations of 2 pg/l or greater (parts per billion; to-tal recoverable basis in 0.45~ filtered samples) should be considered hazardous to thehealth a-rid long-term survival of fish and wildlife populations due to the high potentialfor food-chain bioaccumulation, dietary toxicity, and reproductive effects. In somecases, ultratrace amounts of dissolved and particulate organic selenium may lead tobioaccumulation and toxicity even when total waterborne concentrations are less thanl&l.

Food-chain organisms such as zooplankton, benthic invertebrates, and certainforage fishes can accumulate up to 30 ,ug/g dry weight selenium (some taxa up to 370,ug/g) with no apparent effect on survival or reproduction. However, the dietary toxi-city threshold for fish and wildlife is only 3 pg/g; these food organisms would supplya toxic dose of selenium while being unaffected themselves. Because of this, food-chain organisms containing 3 pg/g (parts per million) dry weight or more should beviewed as potentially lethal to fish and aquatic birds that consume them.

Biological effects thresholds (dry weight) for the health and reproductive successof freshwater and anadromous fish are: whole-body = 4pg/g; skeletal muscle = 8pg/g; liver = 12pg/g; ovaries and eggs = lOpg/g. Effects thresholds for aquatic birdsare: liver = 10 ,ug/g; eggs = 3pg/g. The most precise way to evaluate potential repro-ductive impacts to adult fish and aquatic bird populations is to measure selenium con-centrations in gravid ovaries and eggs. This single measure integrates waterborne anddietary exposure, and allows an evaluation based on the most sensitive biological end-point. If natural resource managers obtain measurements of selenium in water, food-chain organisms, and fish and wildlife tissues, and apply the preceeding guidelinesand assessment protocol, they can accurately determine the overall selenium statusand health of aquatic ecosystems.

CONCLUSIONS

There have been several cases of selenium contamination in the United Statesthat have resulted in severe poisoning of fish and wildlife. Research studies of theseepisodes have generated a data base that clearly illustrates the environmental implica-tions of excessive selenium and the associated threats to public health. Moreover, thedata base has been, and continues to be, used to benefit future generations of wildlifeand humans by providing guidance on what the acceptable and unacceptable concen-trations of selenium are from waterborne and dietary routes of exposure. It is hopedthat this information will be useful for developing countries and nations around theworld because they too may soon have to deal with many of the same problems. Thetwo faces of selenium-required nutrient and potent toxin-make it a paricularly impor-tant trace element in the health of both animals and man. Because of this paradox,environmental selenium in relation to agriculture, fisheries, and wildlife will continueto raise important land and water management issues for decades to come. If these is-sues are dealt with in a timely, prudent manner guided by the environmental seleniumdata base, adverse impacts to natural resources and public health can be avoided.

430 A. DENNIS LEMLY

-_

Hazard fromaccumulation inplanktonic food-chain and dietarytoxicity to fishand aquatic birds

Hazard fromaccumulation inbenthic food-chain and dietarytoxicity to fishand aquatic birds

Hazard fromdietary toxicityand reproductiveimpairment infish and aquaticbirds

High

Moderate

Minimal

None

Cl l-2 2-3 3-5 z-5

Selenium in Water &t/l)

High -

Moderate -

Low -

Minimal -

None - b

Minimal

None

Cl 1-2 2-3 3 - 4 >4Selenium in sediments @g/g dry weight)

High

Moderate

-1

c2 2-3 3-4 4-5 >5Selenium in macroinvertebrates @g/g dry wt.)

SELENIUM IN THE ENVIRONMENT 431

High -

_ Hazard from_ reproductive Moderate -

impairment

L o w - -/-1.--1

Minimal

None

c3 3-5 5-10 1 O-20 >20Selenium in fish eggs @g/g dry weight)

Hazard fromreproductiveimpairment

Minimal -

None -

c3 3-5 5-12 12-20 >20Selenium in aquatic bird eggs @g/g dry wt.)

Fro. 6. Guidelines for rating hazard and predicting the toxic threat of selenium to fish and aquaticbirds based on selenium concentrations. Numerical scores are assigned to each hazard rating andadded to derive a final hazard estimate for the site (see text for explanation) .

REFERENCES

Allen, G. T. and Wilson, R. M. (1990). Selenium in the aquatic environment of Quivira National WildlifeRefuge. Prairie Naturalist 22, 129-135

American Petroleum Institute ( 1978). Analysis of Refinery Wastewaters for the EPA Priority Pollutants.Publication 4296. American Petroleum Institute, Washington, DC.

Andren, A. W. , Klein, D. H., and TaImi, Y. (1975). Se1 enium in coal-fired steam plant emissions. En-viron Sci Technol 9 , 8 5 6 - 8 5 8 .

Braunstein, H. M. , Copenhaver, E. D. , and Pfuderer, H. A. (1981). Environmentul, Health, andControl Aspects of Coal Conversion-An Information Overview, Vol 2. Ann Arbor Science Publishers,Inc., Ann Arbor, MI.

Butler, D. L., Wright, W. G., Hahn, D. A., Krueger, R. P., and Osmundson, B. C. (1994). Physi-

432 A. DENNIS LEMLY

cal. chemical, and biological data for detailed study of irrigation dramage in the Uncompahgre Prolect areaand m the Grand Valley, West-Central Colorado, 1991-92. Open Fde Report 94.110, U.S. GeologicalSurvey, Denver, CO.

Cherry, D. S. and Guthrie, R. K. (1977). Toxic metals m surface waters from coal ash. Wurer Resour.Hlrll 13, 1227.1236.

Clark. P. J. , Zingaro, R. A. , Irgolic, K. J. , and McCinley, A. N. ( 1 9 8 0 ) . Arsemc and selemum inTexas l ignite. Internut, J. Envwon Anulytzcal Chem 7, 2 9 5 3 1 4 .

Clark, D. R. , Jr. ( 1987). Selemum accumulation in mammals exposed to contaminated California irrigationdramwater. Sri. Total Envrron 66, 147-168.

Copeland, R. ( 1970). Selenium: The unknown pollutant. Limnos 3, 7-9.Cowser, K. E. and Richmond, C. R. (1980). Synthetic Fossil Fuel Technology: Potential Health and En-

vzronmental Effects. Ann Arbor Science Publishers, Ann Arbor, MI.Coyle, J.J., Buckler, D. R., Ingersoll, C. G., Fairchild, J. F., and May, T. W. (1993). Effect of di-

etary selenium on the reproductive success of bluegill sunfish ( Lepomts macrochtrus) Environ T&cd.Chem. 12, 551-565.

Cumbie, P. M. and Van Horn S. L. ( 1978). Selenium accumulation associated with fish mortality and re-productive failure. Proc. Annual Conf. Southeast. Assoc. Fish Wildlife Agencies 32, 612-624.

Davis, E.C. and Boegly. W. J., Jr. (1981). Coal pile leachate quality. .I. Environ. Engineer. Div.,Anzer. Sot. Civil Engineers 107. 399-417.

Deverel, S. J, , Fio, J. L. , and Gilliom, R. J. ( 1989). Selenium in tile drain water. Water-Resources In-wstigattons Report 88.4186. U.S. Geological Survey, Sacramento, CA.

Draize, J. H. and Beath, 0. A. (1935). Observations on the pathology of blind staggers and alkali disease.Amer. Veterinary Med. Assoc. _I. 86, 753-763.

Ellis, M. M. , Motley, H. L. , Ellis, M. D. , and Jones, R, 0. f 1937). Selenium poisoning m fishes.Proc. Sot. Expertment. Biol. M e d . 3 6 , 5 1 9 - 5 2 2 .

Evans, D. W., Wiener, J. G., and Horton, J. H. ( 1980). T race element inputs from a coal burningpower plant to adjacent terrestrial and aquatic environments. Air Pollut Control Assoc. /. 30, , 567-573.

Fan, A. M., Book, S. A., Neutra, R. R., and Epstein, D. M. (1988). Selenium and human health im-plications in California’s San Joaquin Valley. _I. Tozicol Environ. Health 23, 539-559.

Fruchter, J. S. and Petersen, M. R. (1979). Environmental characterization of products and effluents fromcoal conversion processes. In Analytical Methods for Coal and Coal Products, Vol. III. , pp. 247-275.Academic Press: NY.

Fuji, R. ( 1988). Water-quality and sediment-chemistry data of drain water and evaporation ponds from Tu-lare Lake Drainage District, Kings County, California, March 1985-March 1986. Open File Report 87-700. U.S. Geological Survey, Sacramento, CA.

Garrett, G. P. and Inmann, C. R. (1984). Selenium-induced changes in fish populations of a heated reser-voir. Proc. Annu. Conf. Southeast. Assoc. Fish Wildife Agencies 38, 291-301.

Gillespie, R. B. and Baumann, P. C. (1986). Effects of high tissue concentrations of selenium on reproduc-tion by bluegills. Trans. Am. Fish. Sot. 115, 208-213.

Harris, T. (1991). Death in the Marsh. Island Press, Covelo Ca.Hartley, W. J. and Grant, A. B. (1961). A review of selenium responsive diseases of New Zealand live-

stock. Federation Proc. 29, 679-688.Heaton, R. C., Williams, J. M. , Bertino, J. P., Wangen, L. E., Nyitray, A. M., Jones, M. M.,

Wanek, P. L., and Wagner, P. ( 1982). Leaching behaviors of high-sulfur coal wastes from two Ap-palachian coal preparation plants. Publication LA-9356-MS. U. S. Department of Energy, Los AlamosNational Laboratory, Los Alamos, NM.

H&z, G. H., Hoffman, D. J., Krynitsky, A. J., and Weller, D. M. G. (1987). Reproduction in mal-lards fed selenium. Environ. Toxic01 Chem 6, 423-433.

Heinz, G. H., Hoffman, D. J., and Gold, L. G. (1989). Impaired reproduction of mallards fed an organ-ic form of selenium. J. Wildlife Manage. 53. 418-428.

Hoffman, D. J. and Heinz, G. H. (1988). Embryotoxic and teratogenic effects of selenium in the diet ofmallards. .I. Toxicol Environ. Health 24. 477-490.

Hoffman, D. J., Ohlendorf, H. M., and Aldrich, T. W. (1988). Selenium teratogenesis in natural popu-lations of aquatic birds in central California. Arch. Environ. Contam Toxicol 17, 519-525.

Hoffman. S. E., Williams, N. J., and Herson, A. I. (1986). The Kesterson story. In Proceedings,First Biennial Conference : Is Current Technology the Answer, pp. 1-9. National Water Supply Improve-ment Association, Springfield, VA.

Jones, D. R. (1990). Batch leaching studies of rundle oil shale. _I. Environ. Qunl. 19, 408-413.Kaakinen, J . W., Jorden, R. M., Lawasani, M. H., and West, R. E. (1975). Trace element behavior

in coal-fired power plant. Environ. .%i Technol 9, 862-869.

SELENIUM IN THE ENVIRONMENT 433

Karlson. U. and Frankenberger. W. T. , Jr. (1989). Accelerated rates of selemum volatilization from Cali-forma soils. Soil Sri. Sm. .-\m J. 53, 749.753.

Karlson. U. and Frankenberger. W. T. , Jr. (1990). Volatilization of selemum from agricultural evapora-tion pond sediments. Scl Tot& Environ 92, 41-54.

Klem, D. H., Andren, A. W., and Bolton. N. E . ( 1 9 7 5 ) . T race element discharges from coal combus-tion-for power production. Water, Azr. Sal Pollut. 5. 71-77.

Kubota,-J. ( 1980). Regional distribution of trace element problems in North America. In Applied SodTrace Elements, pp. 441-466. John Wiley and Sons. New York, NY.

Lemiy, A. D. ( 198Sa). Ecological basis for regulating aquatic emissions from the power industry: The casewith selenium. Reg. Toncol. Pharmacol. 5, 465.486.

Lemly, A. D. ( 1985b). Toxicology of selemum in a freshwater reservoir: Implications for environmentalhazard evaluation and safety. Ecotoricol. Envtron Safety 10, 314-338.

Lemly, A. D. ( 1993a). Guidelines for evaluating selenium data from aquatic monitoring and assessmentstudies. Environ. Monitor. Assess. 28, 83-100.

Lemly, A. D. ( 1993b). Teratogenic effects of selenium in natural populations of freshwater fish. Ecotori-col. Envtron. safety 26, 181-204.

Lemly, A. D. ( 1993~). Metabolic stress during winter increases the toxicity of selenium to fish. AquaticTortcol. 27, 133-158.

Lemly, A. D. (1994a). Agriculture and wildlife: Ecological implications of subsurface irrigation drainage.J. Arid Environ. 28, 85-94.

Lemly, A. D. (1994b). Irrigated agriculture and freshwater wetlands: A struggle for coexistence in thewestern United States. Wetlands Em1 Manage. 3, 3-15.

Lemly, A. D. ( 1995). A protocol for aquatic hazard assessment of selenium. Ecotmicol Environ. Safety32, 280-288.

LemIy, A. D. ( 1996a). Winter Stress Syndrome: An important consideration for hazard assessment ofaquatic pollutants. Ecotoricol Environ. Safety 34, 223-227.

Lemly, A. D. ( 1996b). Assessing the toxic threat of selemum to fish and aquatic birds. Environ. Mont-tar. .4ssess. 43, 19-35.

Lemly, A. D. ( 1997a). Ecosystem recovery following selenium contamination in a freshwater reservoir.Ecotoricol Environ. Safety. 36, 275-281.

Lemly, A. D. ( 1997b). Evaluation of the hazard quotient method for risk assessment of selenium. Ecotosi-col. Environ. Safety. 35, 156-162.

LernIy, A. D. (1997~). Role of season in aquatic hazard assessment. Environ. Monitor. Assess. 45, 89-98.

Lemly, A. D. (1997d). Wastewater discharges may be most hazardous to fish during winter. EnvtronPollut 93, 169-174.

LemIy, A. D. and Smith, G. J. (1987). Aquatic cycling of selenium: Implications for fish and wildlife.Fish and Wildlife Leaflet 12. U. S. Department of the Interior, Fish and Wildlife Service, Washington,DC.

Lemly, A. D., I+tger, S. E., and Nelson, M. K. ( 1993). Sources and impacts of irrigation drainwatercontaminants in arid wetlands. Envrron Tori&. Chem 12, 226.5-2279,

Levander, 0. A. (1987). A global view of human selenium nutrition. Ann. Rev. Nutr. 7, 227.250.Longnecker, M. P. , Taylor, P. R., Levander, 0. A., Howe, S. M., Veillon, C., Mcadam, P. A.,

Patterson, K. Y., Holden, J. M., Stampfer, M. J., Morris, J. S., and Willett, W. C. (1991). Se-lenium in diet, blood, and toenails in relation to human health in a seleniferous area. Am _I. Clin Nu-tr. 53, 1288-1294.

Marsha& E. (1985). Selenium poisons refuge, California politics. Science 229. 144-146.Moore, S. B, Winckel, J., Detwiler, S. J., K l a s i n g , S. A., GuaI, P. A., Kanim, N. R., Kesser, B .

E., Debevec. A. B., BeaxIsley, K., and Puckett, L. K. (1990). Fish and WiLdlife Resources andAgricultural Drainage in the Sun Jmquin Valley California. San Joaquin Valley Drainage Program,Sacramento, CA.

Murtha, M. J., Burnet, G., and Harnby, N. (1983). Power plant fly ash disposal and utilization. Envi-ran. Progress 2. 193-198.

National Research Councii ( 1980a). Energy and the Fate of Ecosystems. National Academy of Sciences,Washington, DC.

National Research Council. ( 1980b). Recommended Dietary Allouatces (9* edition ) National Academy ofSciences, Washington, DC.

North Carolina Department of Natural Resources and Community Development ( 1986). North Carolina waterquality standards documentation: The freshwater chemistry and toxicity of selenium with an emphasis onits effects in North Carolina. Technical Report 86-02. Division of Environmental Management, NorthCarolina Department of Natural Resources and Community Development, Raleigh, NC.

434 A. DENNIS LEMLY

Nystr~m, R. R. and P~sr. G. ( 1982). Chronic effects of ammonia-srripped 011 shale retort water on fishes,birds, and mammals. &clf. Envzron. Contam. Tozrrol. 28. 271-276.

Ohlendorf. H. M., Hoffman, D. _I.. Saikl. .M. K., and Aldrich, T. W. (1986). Embryomc morrallryand abnormalities of aquatlc birds: .4pparent Impacts of selemum from irngatlon drainwater. Sci TotalEnw-on 52. 49.63.

Ohlendorf, H. M., Hothem. R. L., Aldrich, T. W.,-_

and Krymtskl, A. J. ( 1987). Selenium contami-nation of the Grasslands, a major California waterfowl area. Sci Total Environ 66, 169.183.

Ohlendorf, H. M., Hothem, R., and Aldrich. T. W. (1988a). Bloaccumulation of selenium by snakesand frogs in the San Joaquin Valley, Caiifornia. Copecu 1988, 704-710.

Ohlendorf, H. M. , Kilness, A. W. , S’mI mans, J. L., Stroud. R. K., Hoffman, D. J., and Moore, J.F. ( 1988b). Selenium toxicosis in wild aquatx birds. /, Toricol. Environ. Health 24, 67-92.

Pillay, K. K. S., Thomas. C. C., Jr., and Kaminski. J. W. (1969). Neutron activation analysts of theselenium content of fossil fuels. .Vuclear Applications Tech&. 7, 478.483,

Popkin, R. (1986). Kesterson: Nonpoint nightmare. EPA J. 12. 13-14.Presser, T. S. (1994). The “Kesterson Effect”. Environ. Manage. 18, 437.454.Presser, T. S. and Barnes. I. ( 1985). Dissolved constituents including selenium in waters in the vicinity of

Kesterson National Wildlife Refuge and the West Grassland, Fresno and Merced counties, California.Water-Resources Investigations Report 85-4220. U. S. Geological Survey, Menlo Park, CA.

Presser, T. S. and Ohlendorf, H. M. ( 1987). Biogeochemical cycling of selenium in the San Joaquin Val-ley, California, USA. Enwron Manage. 11, 805-821.

Presser, T. S., Swain, W. C., Tidball, R. R., and Severson, R. C. (1990). Geologic sources, mobi-lization, and transport of selenium from the California coast ranges to the western San Joaquin Valley: Areconnaissance study. Water-Resources Investigations Report 90-4070. U. S. Geological Survey, MenloPark, CA.

Presser, T. S. , Sylvester, M. A. , and Low, W. H. ( 1994). Bioaccumulation of selenium from naturalsources in the western states and its potential consequences. Environ. Manage. 18, 423-436.

Rosenfeld, I. and Beath, 0. A. ( 1946). Pathology of selenium poisoning. Wyoming Agri ExperimentSta Bull. 275, l-27.

Saiki, M. K. (1986a). Concentrations of selenium in aquatic food-chain organisms and fish exposed to agri-cultural tile drainage water. In Selenium and Agricultural Drainage: Implications for San FranciscoBay and the California Environment, Proceedings of the Second Selenium Symposium, pp. 25-33, TheBay Institute of San Francisco, Tiburon, CA.

Saiki, M. K. (1986b). A field example of selenium contamination in an aquatic food-chain. In First Annu-al Environmental Symposium : Selenium in the Environment, pp. 67-76. California State University,Fresno. CA.

Saiki, M. K. and Lowe, T. P. ( 1987). Selenium in aquatic organisms from subsurface agricultural drainagewater, San Joaquin Valley, California. Arch Environ. Contam. Toricol. 16, 657.670.

Saiki, M. K. Jennings, M. R., and Hamilton, S. J. (1991). P re lminary assessment of the effects of sele-I’nium in agricultural drainage on fish in the San Joaquin Valley. In The Economtcs and Management ofWater and Drainage in Agriculture, pp. 369-385. Kluwer Academic Publishers. Boston MA.

Saiki, M. K., Jennings, M. R., and Wiedmeyer, R. H. ( 1992 1. Toxicity of agricultural subsurfacedrainwater from the San Joaquin Valley, California. to juvenile chinook salmon and striped bass. Trans.Amer. Fish. Sot. 121, 78-93.

Schlinger, W. G. and Richter, G. N. ( 1980). An evaluation of the Texaco coal gasification process. InProceedings of the First International Gas Research Conference, pp. 150-168. University of ChicagoPress. Chicago, IL.

Severson, R. C. , Fisher, S. E., Jr., and Gough, L. P. (1991). Proceedings of the 1990 Billings LandReclamation Symposium on selenium in arid and semiarid environments, western United States. Circular1064. U. S. Geological Survey, Denver, CO.

Sorensen, E. M. B. (1986). The effects of selenium on freshwater teleosts. In Reviews in EnvironmentalToxicology, 2., pp. 59-116. Elsevier, New York, NY.

Stephens, D. W., Waddell, B., Peltz, L. A., and Miller, J. B. (1992). Detailed study of selenium andselected elements in water, bottom sediment, and biota associated with irrigation drainage in the MiddleGreen River Basin, Utah, 1988-90. Water-Resources Inwstigations Report 92-4084. U. S. GeologicalSurvey, Salt Lake City, UT.

Summers, J. B. and Anderson, S. S. ( 1986). Toxic Substances in Agricultural Water Supply andDrainage : Defining the Problems. U.S. Committee on Irrigation and Drainage, Denver, CO.

Sylvester, M. A., Deason, J. P., Feltz. H. R., and Engberg, R. A. (1991). Preliminary results of theDepartment of the Interior’ s irrigation drainwater studies. In Proceedings of the 1990 Billings LandReclamation Symposium on Selenium in Arid and Semiarid Environments, Western United States, pp115-122. U.S. Geological Survey, Denver, CO.

SELENIUM IN THE ENVIRONMENT 435

Tanji. K., Lauchli, A., and Meyer, J. (1986). Selenium in the San Joaquin Valley. Environment 28. 6-11, 34-39.

U.S. Bureau of Reclamation ( 1981). Water and Power Resource Semites Project Da&z. U. S. Bureau ofReclamation, Washington, DC.

U.S. EPA (U. S. Environmental Protection Agency) (1980). Oil shale symposium: Sampling, analysis,and aualitv assurance. Techmcal Report EPA-600-g/80/022. U.S. EPA. Environmental Research Lab-. _oratory, Cincinnati, OH.

U.S. EPA (U. S. Environmental Protection Agency) (1984). Health effects assessment for selenium (andcompounds). Pu6lirotion PB-86-13-4699. USEPA. Washington, DC.

U. S. EPA (U. S. Environmental Protection Agency) ( 1987). Ambient water quality criteria for selenium- 1987. Technical Report EPA-440/S-87-006. Office of Water Regulations and Standards, U. S.EPA, Washington, DC.

U. S. Fish and Wildlife Service (1986). Prefiminary Surzq of Contaminant Issues of Concern on NationnlWildlife Refuges. Division of Refuge Management, U.S. Fish and Wildlife Service, Washington, DC.

Yang, C., Wang, S., Zhou. R., and Sun, S. (1983). Endemic selenium intoxication of humans in Chi-na. Amer. J. Clin. Nutr. 37. 872-881.

Yang, G., Zhou, R., and Yin, S. (1989a). Studies of safe maximal daily dietary selenium intake in a se-leniferous area in China. I. Selenium intake and tissue levels of the inhabitants. 1. Truce Elm,. Elec-rrolytes He&h Dis. 3, 77-87.

Yang, G. , Yin, S. , and Zhou, R. (1989b). Studies of safe maximal daily dietary selenium intake in a se-leniferous area in China. Part II. Relation between selenium intake and the manifestations of clinical signsand certain biochemical alterations. .I. Trace Elem . Elactrolyfes Health Dis. 3. 123-130.

Zahm. G. R. (1986). Keaterson Reservoir and Katenon National Wildlife Refuge: History, current prob-lems. and management alternatives. Trans. N. Amer. Wildlife Nat. Resow. Gmf. 51. 324-329.

(Received October 4, 1996 Accept& December 19, 1996)